simulateData: Simulate Data for Multi-Species N-Mixture Models

In MultiNMix: Multi-Species N-Mixture (MNM) Models with 'nimble'

Simulate Data for Multi-Species N-Mixture Models

Description

Simulates multi-species correlated abundance data for various Multi-Species N-Mixture (MNM) model types, including standard MNM, Hurdle, AR (autoregressive), and HurdleAR models.

Usage

simulateData(
  model = "MNM",
  R = 10,
  S = 2,
  T = 5,
  prob = "all",
  abundance = "small",
  K = 4,
  theta = 0.5
)

Arguments

model	Character. Specifies the model type. Options are "MNM", "Hurdle", "AR", "HurdleAR". Default is "MNM".
R	Integer. Number of sites. Default is 10.
S	Integer. Number of species. Default is 2.
T	Integer. Number of replicates. Default is 5.
prob	Character. Specifies the range of detection probabilities: "small": Detection probabilities < 0.4. "large": Detection probabilities > 0.5. "all": Detection probabilities between 0.01 and 0.99 (default).
abundance	Character. Specifies the abundance size: "small": Latent species abundance between 0 and 50. "large": Latent species abundance between 0 and 700. Default is "small".
K	Integer. Number of time points (used for AR models). Default is 4.
theta	Numeric. Probability of zero-inflation (used for hurdle models). Default is 0.5.

Details

This function generates abundance data for multi-species N-mixture models under different configurations:

• MNM: Standard multi-species N-mixture model.

• Hurdle: Includes a hurdle component to model zero-inflated data.

• AR: Includes an autoregressive (AR) component for temporal dependencies.

• HurdleAR: Combines hurdle and AR components for zero-inflation and temporal dependencies. The output includes observed and true abundances, detection probabilities, latent variables, and covariance information for the random effects.

Value

A list containing:

• Y: Array of observed abundances.

• N: Array of true abundances.

• p: Array of detection probabilities.

• Sigma: Covariance matrix for the multivariate normal variable a.

• mu: Mean vector for the multivariate normal variable a.

• lambda: Latent abundance rate parameter.

• correlation: Correlation matrix derived from Sigma.

• R, T, S, K: Number of sites, sampling occasions, species, and time points.

• Additional elements depending on the model type:

  • phi: Autoregression parameter (AR and HurdleAR models).

  • muPhi: Mean of the autoregressive parameter (AR and HurdleAR models).

  • varPhi: Variance of the autoregressive parameter (AR and HurdleAR models).

  • zeros: Matrix of zero-indicators for hurdle models.

  • theta: Zero-inflation parameter for hurdle models.

See Also

• simulateData_MNM: Helper function for simulating standard MNM data.

• simulateData_Hurdle: Helper function for simulating hurdle MNM data.

• simulateData_AR: Helper function for simulating AR MNM data.

• simulateData_Hurdle_AR: Helper function for simulating hurdle AR MNM data.

Examples

# Simulate data for a standard MNM model
data <- simulateData(model = "MNM", R = 10, S = 3, T = 5, prob = "all",
abundance = "small")

# Simulate data for a hurdle model
data <- simulateData(model = "Hurdle", R = 10, S = 3, T = 5, prob = "large",
abundance = "large", theta = 0.3)

# Simulate data for an autoregressive model
data <- simulateData(model = "AR", R = 10, S = 2, T = 5, K = 4, prob = "small",
abundance = "small")

# Simulate data for a hurdle autoregressive model
data <- simulateData(model = "HurdleAR", R = 10, S = 3, T = 5, K = 4, prob = "all",
abundance = "large", theta = 0.5)

MultiNMix documentation built on April 3, 2025, 6:10 p.m.